The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes user-provided interview transcripts and descriptions from the $ARGUMENTS variable without using boundary markers or sanitization.
Ingestion points: User-provided interview transcripts and descriptions via $ARGUMENTS in SKILL.md.
Boundary markers: Absent. The instructions do not use delimiters (like triple backticks or XML tags) or explicit instructions to ignore potentially malicious commands embedded in the transcript.
Capability inventory: The skill utilizes Write, Edit, and Bash tools to modify files such as search/context/interview-history.yaml and search/playbook.yaml.
Sanitization: No sanitization or validation of the user input is performed before it is used to generate content for file updates.
[COMMAND_EXECUTION]: The skill uses powerful tools including Bash, Write, and Edit. While these are restricted to the skill's functional scope (managing interview history and playbooks), their use in a workflow that handles untrusted user data creates a potential attack surface if an indirect injection occurs.

interview-debrief